www.nature.com/scientificreports OPEN Four novel algal virus genomes discovered from Yellowstone Lake metagenomes Received: 09 March 2015 1 1,2 3 1,4 1,4 1,5 Accepted: 17 September 2015 Weijia Zhang , Jinglie Zhou , Taigang Liu , Yongxin Yu , Yingjie Pan , Shuling Yan 1,4 Published: 13 October 2015 & Yongjie Wang Phycodnaviruses are algae-infecting large dsDNA viruses that are widely distributed in aquatic environments. Here, partial genomic sequences of four novel algal viruses were assembled from a Yellowstone Lake metagenomic data set. Genomic analyses revealed that three Yellowstone Lake phycodnaviruses (YSLPVs) had genome lengths of 178,262 bp, 171,045 bp, and 171,454 bp, respectively, and were phylogenetically closely related to prasinoviruses (Phycodnaviridae). The fourth (YSLGV), with a genome length of 73,689 bp, was related to group III in the extended family Mimiviridae comprising Organic Lake phycodnaviruses and Phaeocystis globosa virus 16 T (OLPG). A pair of inverted terminal repeats was detected in YSLPV1, suggesting that its genome is nearly complete. Interestingly, these four putative YSL giant viruses also bear some genetic similarities to Yellowstone Lake virophages (YSLVs). For example, they share nine non-redundant homologous genes, including ribonucleotide reductase small subunit (a gene conserved in nucleo-cytoplasmic large DNA viruses) and Organic Lake virophage OLV2 (conserved in the majority of YSLVs). Additionally, putative multidrug resistance genes (emrE) were found in YSLPV1 and YSLPV2 but not in other viruses. Phylogenetic trees of emrE grouped YSLPVs with algae, suggesting that horizontal gene transfer occurred between giant viruses and their potential algal hosts. Phytoplankton (microalgae), based on conservative estimates of more than 100,000 species1, is abundant in the sea. These algae form the base of the marine food web as their photosynthetic activities provide carbon sources and energy for life in the marine ecosystem and they regulate numerous aspects of the global environment1. Marine algae-infecting viruses, particularly phycodnaviruses, are important for controlling the composition of planktonic communities2. The phycodnaviruses are a genetically diverse, morphologically similar group of double-stranded DNA viruses that infect eukaryotic algae and presently contain six genera3. Members of the Coccolithovirus, Phaeovirus, Prasinovirus, Prymnesiovirus and Raphidovirus genera infect marine algae, while chlorovi- ruses (Chlorovirus) infect freshwater algae1,4. The genomes of phycodnaviruses are generally smaller (160 to 560 kb) than those of mimiviruses belonging to Lineage A of Group I, which typically have genomes larger than 1 Mb. Thus far, complete genomes of phycodnaviruses, such as Ostreococcus viruses (OtV1, OtV5)5,6, Micromonas sp. RCC1109 virus MpV17 and five chloroviruses8–10, have been well characterized. The Phycodnaviridae, together with six other giant virus families, were defined as nucleo-cytoplasmic large DNA viruses (NCLDVs)11 and were proposed to be reclassified into a new order Megavirales due to the following shared features12: i) giant viral particles with capsid diameters > 150 nm and genome sizes 1College of Food Science and Technology, Shanghai Ocean University, Shanghai, China. 2Department of Biological Sciences, Auburn University, Auburn, AL, USA. 3College of Information Technology, Shanghai Ocean University, Shanghai, China. 4Laboratory of Quality and Safety Risk Assessment for Aquatic Products on Storage & Preservation, Ministry of Agriculture, Shanghai, China. 5Institute of Biochemistry and Molecular Cell Biology, University of Göttingen, Göttingen, Germany. Correspondence and requests for materials should be addressed to Y.W. (email: [email protected]) SCIENTIFIC REPORTS | 5:15131 | DOI: 10.1038/srep15131 1 www.nature.com/scientificreports/ > 100 kb; ii) the presence of nine class I core genes in all seven families or 47 NCLDV conserved genes in one or two families13; iii) potential for infection with virophages. Interestingly, virophages were found to be associated with giant viruses since the first virophage Sputnik was isolated in association with a mamavirus, a relative of mimivirus, in a water-cooling tower in Paris14. Virophages have circular double-stranded DNA genomes of 18–30 kb which encode more than 20 genes. Recently, Zhou et al.15,16 observed extensive genetic diversity of virophages in Yellowstone Lake metagenomic datasets and have assembled seven complete virophage genomes (Yellowstone Lake virophages, YSLVs). In this study, to provide insight into the diversity of giant viruses in Yellowstone Lake, we assembled giant viral genomes from the same metagenomic datasets in which YSLVs were discovered15,16. Based on comparative genomic and phylogenetic analyses, four novel giant viruses detected in YSL appeared to infect algae, and horizontal gene transfer was observed among giant viruses, YSLVs, and their potential algae hosts. Material and Methods Sequence assembly. Sequence assembly was performed as previously described by Zhou et al.15,16. Briefly, the Yellowstone Lake metagenomics dataset17 was downloaded from the CAMERA 2.0 Portal and assembled de novo using Newbler v2.6 (Roche). Contigs derived from the assembly were con- structed as a local database in order to perform tBLASTx searches for viral major capsid protein (MCP)- related sequences. Since YSLVs are most closely related to Organic Lake virophage (OLV)15,16, and OLV is thought to be the viral parasite of Organic Lake phycodnaviruses (OLPVs)15,18, the potential giant viral hosts of YSLVs may be closely related to OLPVs. Therefore, MCPs of OLPVs were used as refer- ence sequences to search (tblastx, E-value < 10−5) for homologous sequences in the constructed contig database described above. The OLPV MCP-related contigs over 10 kb in length with good quality (low E-value and high identity) were re-assembled using the Yellowstone Lake metagenomic dataset until the assembled sequences no longer extended. All sequence assemblies (with a minimum overlap length of 25 bp and minimum overlap identity of 95%) were performed using GeneiousPro15. Assembly check. To further validate the assembled consensus sequences, duplicate reads were first removed from the data sets of scaffold reads with CD-HIT Suite19. Sequence identity cut-offs were set as: 0.97, 0.95 and 0.90. The obtained unique reads were then re-assembled, and the consensus sequences were compared to the original sequences using the data sets without prior removal of duplicates in order to check the quality and accuracy of sequence assembly. Genomic sequence analysis. The prediction and annotation of open reading frames (ORF) was performed as described by Zhou et al. in 201315. Each predicted ORF contained an ATG start codon, and had a minimum size of 150 bp, standard genetic code, and a stop codon. Translated amino acid sequences were used to search (E-value < 10−1) for homologs in NCBI nr database using the BLASTp program. One top hit to virus and/or non-virus was recorded. Functional annotation of ORFs was performed using the InterProScan program (http://www.ebi.ac.uk/Tools/pfa/iprscan/), Conserved Domain search20 on the NCBI server, and HHpred (http://toolkit.tuebingen.mpg.de/hhpred). In addition, 412 proteins of OLPV1, 326 of OLPV2 and 434 of Phaeocystis globosa virus 16 T (PgV- 16 T) were analyzed for orthologous protein clusters shared with YSLGV proteins using the COG algorithm21. All predicted ORFs (E-value < 10−1) were searched against a local database, which is comprised of all predicted proteins of seven YSLVs. Analysis of PgVV (a pro-virophage associated with PgV-16 T) proteins was performed by searching their homolog against an extensive database containing sequences of all the predicted proteins of known virophages. Genomic sequences were aligned using the Mauve program on the Geneious Pro platform (default parameter)22. Repetitive sequences were checked on the softberry website (http://linux1.softberry.com/ berry.phtml? topic = frep&subgroup = repeat&group = programs) with default parameters, and long ter- minal repeats using LTR-Finder23. Phylogenetic analysis. Homologs of MCP, DNA polymerase B family (PolB), poxvirus late tran- scription factor 3 (Pox-VLTF3), topoisomerase II (Topo II), vaccinia virus (VV) A32-like packaging ATPase, ribonucleotide reductase small subunit (RNR2), multidrug resistance protein (emrE), and OLV ORF2 (OLV2) were used to reconstruct phylogenetic trees. Reconstruction was initiated by aligning multiple amino acid sequences using the MUSCLE program24, followed by tree construction using the JTT model with a bootstrap value of 100. Phylogenetic analysis of emrE included sequences from three cellular life domains and was based on Bayesian Inference (parameter set: rate variation: gamma; rate matrix: poisson). All analyses were performed on the Geneious Pro platform. Accession numbers. The genomic sequences of four YSL algal viruses have been deposited in DDBJ under accession numbers LC015646-LC015649 for YSLGV and YSLPV 1–3, respectively. SCIENTIFIC REPORTS | 5:15131 | DOI: 10.1038/srep15131 2 www.nature.com/scientificreports/ Figure 1. Physical maps of partial genomes of YSLPVs and YSLGV. ORFs are indicated in box arrows and are labeled in different colors that represent different functional categories. Repeats are shown in diamond. Numbers outside the genomes indicate nucleotide positions. Viral names, genomic length and G+ C content, and the total number of predicted ORFs are shown